J Oral Pathol Med 1999; 28: 259–63 Copyright C Munksgaard 1999

Printed in Denmark . All rights reserved JOURNAL OF

ISSN 0904-2512

Expression of p53 oncoprotein Rieko Doi1, Itaru Kuratate1,

Eiji Okamoto1, Kazuo Ryoke2 and
Hisao Ito1

increases intratumoral 1
2
First Department of Pathology and
Department of Oral and Maxillofacial
Surgery, Faculty of Medicine,

microvessel formation in
Tottori University, Tottori, Japan

human salivary gland

carcinomas
Doi R, Kuratate I, Okamoto E, Ryoke K, Ito H: Expression of p53 oncoprotein
increases intratumoral microvessel formation in human salivary gland carcino-
mas. J Oral Pathol Med 1999; 28: 259–63. C Munksgaard, 1999.

P53 protein and vascular endothelial growth factor (VEGF) expression, and
mean intratumoral microvessel density (IMVD) were studied by immunohisto-
chemistry in 31 salivary gland carcinomas, consisting of 11 adenoid cystic carci-
nomas (AdCCs), 10 mucoepidermoid carcinomas (MECs), 7 acinic cell carcinomas
(AcCCs), and 3 squamous cell carcinomas (SCCs). Cases with p53 protein in
more than 20% of tumor cells were detected in one AdCC, four MECs, one AcCC,
and two SCCs. Both frequency of p53 and VEGF expression, and mean IMVD,
were higher in the MECs and SCCs than in the AdCCs and AcCCs. Similarly, Key words: immunohistochemistry;
intratumoral microvessels; p53; salivary gland
both VEGF expression and mean IMVD were significantly higher (P∞0.05) in carcinomas; vascular endothelial growth
the eight p53-positive tumors than in the 23 negative tumors. Six cases with survi- factor
val periods less than 5 years showed significantly higher frequency of p53 and
VEGF expression and of mean IMVD than those with longer survival periods. Hisao Ito, First Department of Pathology,
Faculty of Medicine, Tottori University, 86
These results indicate that p53 expression might partly correlate with VEGF Nishi-cho, Yonago, Tottori 683–8503, Japan
expression and mean IMVD, and be a factor in the survival of patients with
salivary gland carcinomas. Accepted for publication March 10, 1999

Angiogenesis is essentially involved in
growth and progression of various
human solid tumors (1). The process of
intratumoral angiogenesis is regulated
by the local balance of various growth
factors or cytokines released by both
tumor cells and normal cells. Epider-
mal growth factor (EGF), basic fibro-
blast growth factors (bFGF) (2), in-
terleukin-8 (IL-8) (3), and vascular en-
dothelial growth factor (VEGF) (4)
have been shown to act as angiogenetic
factors, while transforming growth fac-
tor b (TGF-b) (5), angiostatin, interfer-
on-g (6) and thrombospondin (7) inhib-
it the process.
Vascular endothelial growth factor
(VEGF), also known as vascular perme-
ability factor, is a secreted protein that
may play a pivotal role in tumor-associ-
ated microvascular angiogenesis and
hyperpermeability (8, 9). Of interest is
that mutated p53 gene has been shown to
be a potent stimulant of VEGF (10).
This implies that mutated p53 gene pro-
vides preferential advantage to tumor
proliferation by not only escaping
apoptosis, but also by forming a favor-
able vascular-rich microenvironment.
Recently, it has been shown that ex-
tent of tumor vascularization correlates
with the prognosis in a variety of cancer
patients, including those with breast
(11), bronchial (12), colonic (13) and
gastric carcinomas (14–16). In gastric
carcinomas, for example, higher intra-
tumoral microvessel density (IMVD)
was confirmed in the tumors with
lymph node metastases than in those
without such metastases (14). On the
other hand, T et al. reported
that the IMVD correlated with liver
metastases, but not with lymph node
metastases (15). T et al. found
a good correlation between IMVD and
VEGF expression in intestinal-type gas-
tric cancers (16). Thus, IMVD and clin-
ical relevance have been confirmed, but
the precise mechanism of angiogenesis
has not been fully elucidated, especially
the role of the p53 gene.
Few reports are available on angio-
genesis and its clinical relevance in the
field of oral pathology. Attention has
been focused on oral squamous cell car-
cinomas, in which the clinical and bio-
logical significance of IMVD are con-
260 D et al.
Table 1. Histology and the expression of p53 and angiogeneic activity
Material and methods
Number Mean p53- VEGF¿ ....-
Cases and clinical information
of age [yr] positive positive IMVD¿....
Histological type cases (range) tumors tumors (mean∫SD) Studies were conducted on a total of
Adenoid cystic carcinoma 11 59.0 1 7 17.3∫11.0
31 salivary gland carcinomas, con-
(34∂77) sisting of 11 adenoid cystic carcino-
Mucoepidermoid carcinoma 10 66.6 4 7 42.7∫31.4
mas (AdCCs), 10 mucoepidermoid
(58∂84) carcinomas (MECs), 7 acinic cell car-
Acinic cell carcinoma 7 57.6 1 3 19.5∫17.0
cinomas (AcCCs), and 3 squamous
(35∂79) cell carcinomas (SCCs). All these cases
Squamous cell carcinoma 3 68.7 2 3 45.2∫8.2
were surgically resected at the Depart-
(60∂76) ment of Oral and Maxillofacial
Surgery, Tottori University and its re-
.... VEGF: vascular endothelial growth factor; IMVD: intratumoral microvessel density.
¿ lated Teaching Hospitals, Japan, dur-
ing the period from 1974 to 1994. We
obtained clinical information on age,
troversial (17–20). With this back- VEGF expression, IMVD, and also sex, size, localization, tumor recur-
ground, we attempted to examine the clinical relevance in salivary gland car- rence and disease outcome from the
correlation between p53 protein and cinomas. clinical files of each patient. None of

Fig. 1. A) Histology of a mucoepidermoid carcinoma of minor salivary gland (H&E, ¿340, barΩ50 mm). B) Immunohistochemical staining
of p53 in a mucoepidermoid carcinoma; a few tumor cells show p53 immunoreactivity in their nuclei (¿340, barΩ50 mm). C) Immunohisto-
chemical staining of vascular endothelial growth factor in a mucoepidermoid carcinoma (¿680, barΩ50 mm). D) Immunohistochemical
staining of CD34; intratumoral microvessels are clearly demonstrated (¿340, barΩ50 mm).

the patients had received preoperative
adjuvant therapy (radiotherapy or che-
motherapy). The specimens were all
fixed in 10% buffered formalin and
embedded in paraffin wax. Histologi-
cal classification was made according
to the WHO classification (21). Two
or three representative blocks were se-
lected and serial sections were exam-
ined by immunohistochemistry.
Immunohistochemistry and assessment
of intratumoral microvessel density
Serial sections cut at 4 mm thickness
were stained for p53, CD34 and VEGF.
Immunohistochemical staining was per-
formed by the avidin-biotin-peroxidase
complex technique, following micro-
wave-based treatment for antigen re-
trieval. The antibodies used were the
monoclonal antibodies for p53 (BP53-
12, Novocastra Laboratoriers Ltd,
Newcastle, UK) and CD34 (Histofine,
Nichirei Co, Tokyo, Japan) at a 1:50 di-
lution, and rabbit polyclonal antibody
for VEGF (A-20, Santa Cruz Biotech-
nology, Santa Cruz, CA, USA) at a
1:150 dilution. For negative controls,
non-specific IgG was used as the pri-
mary antibody. Slides were counter-
stained with methyl green.
Cases were considered to be p53- and
VEGF-positive when staining was ob-
served in more than 20% of tumor cells.
The proportion of positive-staining tu-
mor cells was evaluated by two of the
authors (RD and HI), independently,
and then the grade was adjusted. This
classification was based on reported
findings that demonstrated the presence
Fig. 2. A) Histology of an adenoid cystic carcinoma of submandibular gland (H&E, ¿340, barΩ50 mm). B) Immunohistochemical staining of
CD34, which shows only a few intratumoral microvessels (¿40, barΩ50 mm).
P53 and microvessels of salivary carcinomas 261
of p53-positive cells, even in non-neo-
plastic mucosa of the gallbladder, co-
lon, and stomach in a scattered manner
(22–25). However, diffuse or numerous
p53-positive cells have not been re-
ported in non-neoplastic mucosa.
Moreover, some authors have indicated
that a diffuse pattern of p53-positive
cells could be categorized as overexpres-
sion of p53 protein and a marker for
carcinoma (23, 25). The cases with
more than 20% of p53-positive carcino-
ma cells were considered to show a dif-
fuse pattern. Classification of VEGF-
positive cells was followed by that of
p53-positive cells.
For determination of intratumoral
microvessel density (IMVD), the most
vascular-rich five to six areas within the
section were selected and counted under
a light microscope with 100-fold magni-
fication as described by W et al.
(26). The average counts were recorded
as IMVD.
The statistical analysis was per-
formed using Student’s t-test or chi-
squared test. A P value below 0.05 was
considered significant.
Results
Table 1 shows the overall results for
p53, VEGF and CD34 immunohisto-
chemistry. Cases with p53 immunoreac-
tivity in more than 20% of tumor cells
were found in one, four, one, and two
of the 11 AdCCs, 10 MECs, 7 AcCCs
and 3 SCCs, respectively. Similarly,
VEGF immunoreactive tumor cells
were detected in 7, 7, 3 and 3 cases, re-
spectively. Although there was a ten-
dency for the frequency of p53 or
VEGF expression to be higher in the
MECs (Fig. 1a–c) and SCCs than in the
AdCCs (Fig. 2a) and AcCCs, the small
number of cases involved did not de-
monstrate a statistical difference. CD34
immunohistochemistry clearly demon-
strated intratumoral microvessels, and
the density was highest in the SCCs, fol-
lowed by the MECs (Fig. 1d), AcCCs
and AdCCs (Fig. 2b), in the order giv-
en. Intratumoral microvessel density
(IMVD) was significantly higher in the
MECs and SCCs than in the AdCCs
and AcCCs (P∞0.05). The former two
tended to show a higher frequency of
p53 protein (Fig. 1b) and VEGF-posi-
tive tumors (Fig. 1c).
All the cases were divided into two
groups: p53 protein-positive and nega-
tive tumors. Mean IMVD was
48.0∫30.6 in the 8 positive tumors and
22.0∫16.7 in the 23 negative cases, the
value being significantly different
(P∞0.05). In addition, VEGF-positive
tumors were found in 7 (87.5%) of the
former and in 12 (56.5%) of the latter.
The frequency was significantly higher
in the p53-positive cases, implying a
close relationship between p53 protein
and VEGF expression in the tumor
cells. Because of the small number of
cases, these analyses could not be con-
ducted in each histological category, se-
parately, except for the MECs, in which
mean IMVD was 55.3∫41.8 in the four
p53-positive tumors and 34.4∫22.9 in
the six negative tumors – the value be-
ing significantly different (P∞0.05).
Clinicopathological features were
compared with p53 protein and VEGF
262 D et al.
Table 2. Metastases and the expression of p53, IMVD and VEGF
Number
of p53-
Metastases cases positive (%)
Positive 15 6 (40.0)
Lymph nodes 4 1
Lymph nodes and distant sites 3 1
Distant sites 8 4
Negative 16 2 (12.5)
*: P∞0.05
Table 3. Survival period and the expression of p53, IMVD and VEGF
Number p53-
Survival period of cases positive (%)
More than 5 years 23 4 (17.4)
Less than 5 years 6 4 (66.7)
*: P∞0.05
immunohistochemistry, and also with
mean IMVD in the primary lesions.
Table 2 shows the correlation between
tumor metastasis and the parameters
described above. Fifteen cases showed
tumor metastasis in regional lymph
nodes and/or distant organs, including
liver and lung. P53 protein was found
in 6 (40.0%) of the 15 tumors with me-
tastases and in 2 (12.5%) of the 16 tu-
mors without metastases, the value be-
ing significantly higher in the former
(P∞0.05). On the other hand, mean
IMVD and VEGF positivity did not
differ between the two groups. How-
ever, eight tumors with distant organ
metastases showed higher frequency of
VEGF expression (87.5%) than the
other tumors.
The relationship between patient sur-
vival period and the three parameters is
shown in Table 3. The patient’s survival
period was available in 29 cases. P53-
positive tumors were detected in 4
(66.7%) of the 6 patients with a survival
period less than 5 years, and in 4
(17.4%) of the 23 patients who survived
for more than 5 years. Similarly,
VEGF-positive tumors were found in 5
(83.3%) of the former and in 14 (60.9%)
of the latter. Moreover, mean IMVD
was 41.9∫11.9 in the former and
25.7∫25.2 in the latter. Thus, both the
frequency of p53- and VEGF-positive
tumors, and of mean IMVD, were sig-
nificantly higher in the patients with
shorter survival periods than in those
with longer survival periods (P∞0.05).
Discussion
In the present study, we have demon-
strated the correlation between p53 and
VEGF- IMVD
positive (%) (mean∫SD)
10 (66.7) 28.9∫15.9
(25.0) 1 (25.0) 21.7∫15.9
(33.3) 2 (66.7) * * 41.7∫10.9
*
(50.0) 7 (87.5) 27.7∫16.2
10 (62.5) 28.5∫29.7
VEGF- IMVD
positive (%) (mean∫SD)
14 (60.9) 25.7∫25.2
* *
5 (83.3) 41.9∫11.9
VEGF expression, and IMVD, in 31
human salivary gland carcinomas, com-
pared with histological type and
clinicopathological parameters. Both
the frequency of p53 and VEGF expres-
sion, and IMVD values, were higher in
the MECs and SCCs than in the
AdCCs and AcCCs. Although the small
number of the cases did not allow us to
draw a definitive conclusion, the results
might partly reflect the biological be-
havior of the salivary gland carcinomas
of each histological type. However, no
apparent difference has been reported
on the prognosis between MECs and
AcCCs (27–29).
A few reports are available on p53
expression in human salivary gland car-
cinomas, except for carcinoma in pleo-
morphic adenomas (27, 28, 30–33). Our
results were consistent with the reports
of S et al., who found lower fre-
quency of p53 expression in MECs and
absence in AdCCs and AcCCs (28). Re-
cently, a relative higher frequency of
p53 expression has been reported by
G et al. and K̈̈ et al. (27, 31).
The variation in p53 expression is most
apparent in the AcCCs. A lower fre-
quency was reported by S et al.
(0%, 0 of 6 cases) (28), and by I &
N (13%; 3 of 24) (28), while a
higher frequency was noted by G
et al. (80%; 8 of 10)(31) and by Ä̈ et
al. (69%; 18 of 26) (27). This difference
in frequency of p53 expression might be
partly due to the antibody used. More-
over, none of the reports paid attention
to the histological grade of the AcCCs,
which showed a wide spectrum of histo-
logical differentiation (34, 35). S
et al. examined thyroid tumors and
found that the percentage of p53-posi-
tive tumor cells was significantly higher
in poorly differentiated and undif-
ferentiated carcinomas than in papillary
and follicular carcinomas (36). B-
 et al. reported p53 gene mutation in
anaplastic variants, but not in non-ana-
plastic types, of Wilms’ tumors (37).
Thus, it is feasible that p53 expression
correlates with histological differentia-
tion, which awaits further clarification
by analyzing a large number of AcCCs.
As alluded to in the introductory sec-
tion, VEGF expression has been shown
to be regulated by p53 protein. Mutated
p53 gene is a potent stimulant of VEGF
(10). We confirmed that adenovirus-vec-
tor-mediated transfer of wild-type p53
*
gene downregulated VEGF expression
in a human gastric carcinoma cell line,
MKN-1 (38). Results of the present
study also indicated a close correlation
between p53 gene and VEGF expres-
sion. Both frequency of VEGF and
mean IMVD were significantly higher
in the eight p53-positive tumors than in
the 23 negative tumors. Recently, Y-
 et al. clearly demonstrated VEGF
mRNA expression in not only surgical
specimens but also in all eight human
gastric cultured cell lines, regardless of
histological type (39). However, they
could not find a correlation between
p53 gene status and VEGF expression.
These results might indicate that VEGF
expression is regulated by not only the
p53 gene, but also by other factors.
In summary, we have demonstrated
that p53 expression, might at least part-
ly, correlate with VEGF expression and
mean IMVD, and be a clinically useful
factor in determining the prognosis of
patients with salivary gland carcinomas,
irrespective of histological type.
Acknowledgements – We thank T T-
 (Department of Oral Pathology, Hiro-
shima University) and Mr. N I
(First Department of Pathology, Tottori Uni-
versity) for technical advice, and Mr. H-
 M (Red Cross Hospital of Matsue, Ja-
pan) and Mr. S N (Tottori Pre-
fectural Central Hospital, Japan) for
providing materials. This work was support-
ed in part by Grants-in-Aid for Cancer Re-
search from the Ministry of Education, Sci-
ence, Sports and Culture, 2nd term Compre-
hensive 10-year strategy for Cancer Control,
and the Ministry of Health and Welfare of
Japan.
References
1. L DW, C G, K W-J,
G DV, F N. Vascular en-
dothelial growth factor is a secreted an-

giogenic mitogen. Science 1989; 246:
1306–9.
2. O S, I M, T SH,
E M, B C. Neuronal de-
fects and delayed wound healing in mice
lacking fibroblast growth factor 2. Proc
Natl Acad Sci USA 1998; 95: 5672–7.
3. Y A, Y S, K AK, I-
 T, I H. Role of NF-kap-
pa B-mediated interleukin-8 expression
in intraocular neovascularization. Invest
Ophthalmol Vis Sci 1998; 39: 1097–106.
4. S K, I K, O M, et al.
Expression of vascular endothelial
growth factor and its possible relation
with neovascularization in human brain
tumors. Cancer Res 1995; 55: 1189–93.
5. D J JS, V D PJ, V D
V P, B JP. Expression of growth
factors, growth-inhibiting factors, and
their receptors in invasive breast cancer.
II. Correlations with proliferation and
angiogenesis. J Pathol 1998; 184: 53–7.
6. N LF, M G, B O.
The immunobiology of interferon-gam-
ma inducible protein 10 kD (IP-10): a
novel, pleiotropic member of the C-X-C
chemokine superfamily. Cytokine
Growth Factor Rev 1997; 8: 207–19.
7. V OV, L J, B NP. A
human fibrosarcoma inhibits systemic
angiogenesis and the growth of experi-
mental metastases via thrombospondin-
1. Proc Natl Acad Sci USA 1998; 95:
6343–8.
8. B B, B LF, V D W L,
D HF, S DR. Vascular per-
meability factor (vascular endothelial
growth factor) gene is expressed differ-
entially in normal tissues, macrophages,
and tumors. Mol Biol Cell 1992; 3: 211–
20.
9. D V C, E JA, U H,
H K, F N, W LT.
The fms-like tyrosine kinase, a receptor
for vascular endothelial growth factor.
Science 1992; 255: 989–91.
10. K A, W HA, B G,
M D, K W. Mutant p53 po-
tentiates protein kinase C induction of
vascular endothelial growth factor ex-
pression. Oncogene 1994; 9: 963–9.
11. T M, H S, T T, T-
 T. Association of vascular endo-
thelial growth factor expression with tu-
mor angiogenesis and with early relapse
in primary breast cancer. Jpn J Cancer
Res 1994; 85: 1045–9.
12. F EA, R D, M
KM. Neovascularization in hyperplastic,
metaplastic and potentially preneoplas-
tic lesions of the bronchial mucosa.
Virchows Arch 1996; 429: 95–100.
13. T Y, K Y, B CD.
Expression of vascular endothelial
growth factor and its receptor, KDR,
correlates with vascularity, metastasis,
and proliferation of human colon can-
cer. Cancer Res 1995; 55: 3964–8.
P53 and microvessels of salivary carcinomas 263
14. M K, C Y-S, T S, et
al. Tumor angiogenesis and tumor cell
proliferation as prognostic indicators in
gastric carcinoma. Br J Cancer 1995; 72:
319–23.
15. T N, A H, M
M, et al. Extent of tumor vascularization
correlates with prognosis and hematoge-
nous metastasis in gastric carcinomas.
Cancer Res 1996; 56: 2671–6.
16. T Y, B CD, A Y, et
al. Significance of platelet-derived endo-
thelial cell growth factor in the angio-
genesis of human gastric cancer. Clin
Cancer Res 1998; 4: 429–34.
17. G LL, B PW, P ZP,
G JL. Tumor angiogenesis in
T1 oral cavity squamous cell carcinoma:
role in predicting tumor aggressiveness.
Head Neck 1996; 18: 343–6.
18. L DA, T DR, K JD,
W N, C JI. Tumor angiogen-
esis, the p53 antigen, and cervical metas-
tasis in squamous cell carcinoma of the
tongue. Otolaryngol Head Neck Surg
1994; 111: 417–22.
19. P CN, P M, R R,
L JD. The role of angiogenesis
in the spread of oral squamous cell carci-
noma. Br J Oral Maxillofac Surg 1996;
34: 37–41.
20. S S, R A, Y Y,
O K, M T. Immunohis-
tochemical detection of microvessels in
oral squamous cell carcinomas. Jpn J
Oral Maxillofac Surg 1994; 40: 1–6.
21. World Health Organization. Histological
typing of salivary gland tumors: interna-
tional histological classification of tu-
mors. No. 7. 2nd edn. Geneva, Switzer-
land: WHO, 1991.
22. O Y, W H, A Y,
H K. P53 immunostaining
distinguishes malignant from benign
lesions of gallbladder. Pathol Int 1995;
45: 58–65.
23. B IO, M JWR, O
GJA, V B, H SR. Im-
munohistochemistry of altered p53 sup-
pressor gene product in colorectal neo-
plasms. Lab Invest 1993; 68: 234 (abstr).
24. G Y, O M, K N, I H.
Numerical aberration and point muta-
tion of p53 gene in human gastric intesti-
nal metaplasia and well-differentiated
adenocarcinoma: analysis by fluores-
cence in situ hybridization (FISH) and
PCR-SSCP. Int J Cancer 1996; 66: 594–
9.
25. V  B FM, B IP, P M,
O GJ. Detection of p53 overex-
pression in routinely paraffin embedded
tissue of human carcinomas using a nov-
el target unmasking fluid (TUF). Am J
Pathol 1993; 142: 381–5.
26. W N, S JP, W WR,
F J. Tumor angiogenesis and me-
tastasis: correlation in invasive breast car-
cinoma. N Engl J Med 1991; 324: 1–8.
27. K̈̈ VJ, S̈ KJ, K A-
K, S̈ SM. Expression and mut-
ations of p53 in salivary gland tumors. J
Oral Pathol Med 1997; 26: 217–23.
28. S Y, K D, N K, L DP,
V̈̈ K, P̈̈̈ P. Low p53
protein expression in salivary gland tu-
mors compared with lung carcinomas.
Virchows Arch A Pathol Anat 1992; 421:
415–20.
29. L X, T T, W S, S F, W
Z, S F. Cytoplasmic expression
of p53 protein and its morphological
features in salivary gland lesions. J Oral
Pathol Med 1995; 24: 201–5.
30. D H, H H, H Y. C-
myc, ras, p21 and p53 expression in pleo-
morphic adenoma and its malignant
form of the human salivary glands. Acta
Pathol Jpn 1993; 43: 413–422.
31. G O, F A, B S, B
V, G E, A E. P53 onco-
protein expression in parotid gland car-
cinoma is associated with clinical out-
come. Cancer 1995; 75: 2037–44.
32. H HB, K MG, N
C. Salivary duct carcinoma: a highly ag-
gressive salivary gland tumor with over-
expression of c-erbB-2. J Pathol 1994;
172: 35–44.
33. I K, N T. Evaluation of ma-
lignant grade of salivary gland tumors:
studies by cytofluorometric nuclear
DNA analysis, histochemistry for nucle-
olar organizer regions and immuno-
histochemistry for p53. Pathol Int 1994;
44: 287–96.
34. L R, R̈ R, Z A, A
C. Retrospective analysis of 35 patients
with acinic cell carcinoma of the parotid
gland. J Oral Maxillofac Surg 1998; 56:
440–3.
35. T CI, D I, P T,
T J, E G, G P. Clinico-
pathological predictors of recurrence for
acinic cell carcinoma. Clin Otolaryngol
1995; 20: 396–401.
36. S P, C-T J, S-
-S̃ M. Immunohistochemical de-
tection of p53 in differentiated, poorly
differentiated and undifferentiated carci-
nomas of the thyroid. Histopathology
1994; 24: 205–10.
37. B N, B JB, P
J. Clonal expression and attenuated
apoptosis in Wilms’ tumors are associ-
ated with p53 gene mutations. Cancer
Res 1995; 55: 215–9.
38. S T, M T, I M,
K N, I H. Mutation of p53
gene and angiogenesis in human well dif-
ferentiated gastric carcinomas. 70th Jpn
Gastric Cancer Cong 1998, Tokyo, Japan
(abstr).
39. Y S, Y W, K Y, et al.
Expression of vascular endothelial
growth factor in human gastric carcino-
mas. Pathol Int 1998; 48: 499–506.